Smoking cessation is analyzed from the perspective of persistence of changes in cognitions (beliefs, attitudes, and intentions) and actions. Changes in beliefs and attitudes are hypothesized to occur as attempts to reduce cognitive dissonance that is induced by the initial change in smoking behavior. A smoking cessation treatment program comprised of components designed to change smokers' beliefs and attitudes and smoking behavior is demonstrated and evaluated. The evaluation design involves two treatment groups. One receives the full treatment package containing both cognitive and behavior modification components, and the other receives behavior modification treatment only. Behavior modification techniques utilized include self-control, behavior contracting, stimulus control, and covert sensitization. Comparison of the smoking reduction achieved in the two groups provides an assessment of the contribution of belief and attitude components to smoking cessation.